Benchmarking micro finance institutions in India and determinants of their technical efficiency.

Varman, P. Mahendra

Abstract

This study has made a modest attempt to benchmark the best practice MFIs in India following Stochastic Production Frontier Approach. Accordingly Satin Credit Care and IASC have taken the position of Benchmark MFIs. The results from the regression analysis for examining determinants lead to conclusion that the size of MFI, the Legal status and Year of Experience are important in the determination efficiency of MFIs

I. INTRODUCTION

Globalization has brought substantial benefits around the world, but in many developing countries it is contributing to a growing disparity between the rich and the poor. In a country like India, The structure of the economy is dualistic. We can see the Growing Companies, booming stock market and soaring profits, making the rich, richer on the one hand while faltering incomes and wages in the field of agriculture and allied activities is making the poor, poorer on the other. This worsens the access of the poor to the economic opportunities through which they could build up their assets and enhance income in order to come out of poverty cycle. The potentials to avail such economic opportunities mainly depends on the degree of access to financial services. The commercial banking sector does not consider the poor bankable owing mainly to their inability to meet the eligibility criteria, including collateral. Thus the poor in most countries have had no access to formal financial services.

Due to the above mentioned reasons, the rural poor were relying on informal credit channels such as local moneylenders, market vendors, shopkeepers and others including friends and relatives. Credit in the informal system is usually available immediately, when and where required and often without collateral & lengthy documentation formalities, since the lender relies on personal knowledge of the borrower and his surroundings. However the interest rates are not only extremely high, but sanctions often include conditions, verbal or written, which are heavily loaded in favour of the lender and are detrimental to the interest of the borrowers (Nanda, NABARD). (1)

The more rational way to help the poor could be the provision of sustainable economic opportunities at gross-root level especially provision of required financial services at competitive rates to support their investments and viable business activities. India is perhaps the largest emerging market for microfinance. Over the past decade, the Microfinance sector has been growing in India at a fairly steady pace. Though no microfinance institution (MFI) in India has yet reached anywhere near the scale of the well-known Bangladeshi MFIs, the sector in India is characterized by a wide diversity of methodologies and legal forms.

High administrative costs are involved in forming groups and disbursing group based credits and ancillary socioeconomic inputs. For instance Bangladesh Rural Advancement Committee (BRAC), largest NGO in Bangladesh has been providing small loans to a large number of people, who require sustained access to formal credit for generating employment and income. Despite the fact that organizational discipline, skill development and training are likely to promote proper loan use and high loan recovery, it is unlikely that the BRAC could have generated sufficient revenue in the early years of its operation to support these costs. On the other hand, it is unlikely that the rural poor could have borne the full cost of BRAC activities.

Given its high loan recovery rates and low default costs, the cost of BRAC's operations depends to a large extent on the availability and sources of funds for lending and other program activities. (Khandker & Khalily, 1996). Like wise in India too, many MFIs depend on donors for funds at concessionary rates that have been instrumental to the expansion and institutional development of these NGOs and MFIs. Reliance on these funds raises serious questions about the NGO's / MFI's sustainability. Would it remain viable without these subsidized funds? What impact would a non-subsidized operation have on the poor? If subsidy is unavoidable, then how much is required to sustain such activities? And for how long? Can it be sustained and is it worth continuing?

However, Sustainability itself has to be seen in a broader sense than just financial sustainability (Mahajan & Nagasri, 1999).

II. MICRO FINANCE: THE INDIAN SCENARIO

Microfinance emerged as a noble substitute for informal credit and an effective and powerful instrument for poverty reduction among people who are economically active but financially constrained and vulnerable in various countries (Morduch and Haley, 2002). It covers a broad range of financial services including loans, deposits and payment services, and insurance to the poor and low-income households and their microenterprises. Convincing research evidence exists showing significant role of MFIs in improving the lives of the deprived communities in various countries. Persuaded with the potential role of micro financing in alleviating poverty, the South Asian countries especially India & Bangladesh have been actively pursuing the policy of setting up formal network of microfinance institutions. These institutions include NGOs/NBFCs and government sponsored programs.

In India the SHGs through MFIs made a beginning with one of the NABARD funded projects on 'savings and credit management of self help groups' of Mysore Resettlement and Development Agency (MYRADA) in 1986-87. Again in 1988-89, NABARD undertook a survey of 43 NGOs spread over eleven states in India, to study the functioning of SHGs and possibilities of collaboration between the banks and SHGs in the mobilization of rural savings and improving the delivery of credit to the rural poor (Dasgupta, 2001).

The survey results were encouraging, which made NABARD to impress upon Reserve Bank of India (RBI). In response to that RBI, in July 1991, advised the commercial banks (later RRBs and Co-operatives) to extend credit to the SHGs under the pilot project of NABARD, wherein 500 SHGs all over India were covered. Since then SHGs became a regular component of the Indian financial system. Further, in 1994 RBI constituted a working group to formalize the mechanism to larger extent and to review the functioning of NGOs/MFIs and SHGs in expanding and deepening their role in the rural sector. In response to the recommendations RBI advised the bank lending to SHGs through MFIs be considered as an additional segment under priority sector.

The organisations engaged in microfinance activities in India may be categorised as the Wholesalers, NGOs supporting Self Help Group Federations (SHGF) and NGOs directly retailing credit borrowers or groups of borrower. The wholesale agencies which provide bulk funds to the system through NGOs include the National Bank of Agriculture and Rural Development (NABARD), Rashtriya Mahila Kosh-New Delhi and the Friends of Women's World Banking in Ahmedabad. The NGOs that are supporting the SHG Federations include MYRADA in Bangalore, Self-help Women.s Association (SEWA) in Ahmedabad, PRADAN in Tamilnadu and Bihar, ADITHI in Patna, SPARC in Mumbai, and the Association for Sarva Seva Farms (ASSEFA) in Madras, the Small Industries Development Bank of India (SIDBI) and the Tamil Nadu Womens' Development Corporationetc. The NGOs that are directly enhancing credit to the borrowers include SHARE in Hyderabad, ASA in Trichy, RDO Loyalam Bank in Manipur (Tiwari and Fahad, 2004). There are perhaps 150-200 NGOs/MFIs in the field of micro-finance. Currently there are more than 13 million active borrowers in India.

III. MFIS AND PROBLEM OF ITS SUSTAINABILITY

Some leading MFIs, e.g. Grameen Bank in Bangladesh, have created financial modes that serve increasing number of poor. They also lead to repayment rates positively comparable with the performance of many commercial banks. These approaches have helped many MFIs in achieving a reasonable level of sustainability, and have even produced profits without government subsidies and support from donor (Huhne, 1999). Nonetheless, some of the MFIs especially the NGOs are facing serious sustainability problems indicating lapse in their financial procedures, organizational design and governance. Moreover, most of the MFIs do not provide deposit services to their clients. In contrast, Some of the successful MFIs like Grameen Bank in Bangladesh and BancoSol in Bolivia have incorporated the provision of deposit services in their operations. Appropriately, managing the deposit service and micro and small savings help MFIs to reach financial self-sufficiency through generating their own internal flow of funds that in turn reduce their dependency on external sources (Morduch and Haley, 2002). The MFIs exclusively dependent on external sources of funding usually are not sustainable and efficient (Rhyne, 1998).

Scores of studies are found on analyzing the efficiency and its determinants in commercial banking sectors of various countries. The MFIs are also financial institutions with a primary objective of making credit available to that segment of the population which has been ignored by the commercial banking system for not having collateral requirements. The efficient functioning of these MFIs on sustainable basis is important also for persistent financial access of the poor segment of the society. There is dearth of literature regarding efficiency analysis of MFIs in India. However, a few examples for other countries and regions are found in literature such as Nghiem (2004) Nieto, Cinca and Molinero (2004) and Leon (2003) using data from Vietnam, Latin America and Peru, respectively.

Therefore the primary objective of this study is to identify the most efficient/best practice MFI(s) and benchmark it such that it would in turn be helpful to improve functioning of the other MFIs in India. Secondly to identify the factors that improves the level of efficiency of MFIs.

IV. ASSESSING MICRO FINANCE INSTITUTION'S PERFORMANCE

Measuring MFI's performance draws its base from the concepts of stochastic production frontier approach and technical efficiency. While the production possibility frontier shows the maximum amounts of output that can be attained for given quantity of available inputs, technical efficiency characterizes the relationship between observed outcome and the potential outcome that could be obtained from a given level of inputs employed (Farrel, 1957). The measurement of MFI's performance based on these principles seeks to establish the relation between the observed levels of outcome with a given set of inputs and the potential outcome levels that the MFI's could have achieved with the same set of inputs.

As a prelude to estimating the MFIs performance, the concept of 'efficiency' or MFIs performance as explained by Murray and Frenk (1999) and Evan et al. (2001) is illustrated in figure 1.

[FIGURE 1 OMITTED]

The goal of the MFI's performance as reflected by outcomes--Gross Loan Portfolio or Average Loan Balance per Borrower etc., is measured on the vertical axis, while the inputs like No. of Loan officers and Cost per Borrower used to achieve the goal are measured on the horizontal axis. The upper line represents the frontier or the maximum possible output level that could be obtained for a given level of inputs. The lower line depicts the output level that would occur in the absence of the MFI's Inputs.

Assuming a MFI is observed to have achieved (a) + (b) units of output against the maximum possible of (a) + (b) + (c) units, the MFI performance under the scenario will be (b) / (b + c). This indicates what the MFI achieves as compared with its potential, implying that it could have achieved better outcomes with the resources that it has invested. The distance between a MFI's actual level of Output and the frontier is called its 'efficiency'.

V. THE MODEL

In order to assess the performance of MFI's across twenty six sample MFI's in India over the period selected for study (2005 - 2007), the panel data method of estimation was adopted and the empirical model used for estimation can be specified as:

[Y.sub.it] = [[alpha].sub.i] + [X'.sub.it][beta] + [v.sub.it] - [u.sub.i] (1)

Where 'i' indicates individual MFIs and 't' indicates time. [Y.sub.it] denotes the outcome for MFI 'i' in time 't' while [X'.sub.it] is the vector of MFI's inputs, [v.sub.it] represents the random disturbances beyond the control of the MFI, minor omitted variables and measurement errors. The [v.sub.it] are uncorrelated with the independent variables [X'.sub.it] and is normally distributed. The ui represents technical (in)efficiency and correspondingly [u.sub.i] [greater than or equal to] 0, for all 'i'

The ui is assumed to be normally distributed with mean [mu] and [alpha]2u independent of the vit. The technical efficiency ui reflects that the MFIs must lie on or below the frontier, [alpha] + [X'.sub.it][beta] + [v.sub.it]. Although [u.sub.i] is unobserved its existence indicates that the MFIs take into consideration [u.sub.i] while demanding resources for future use.

The technical efficiency is independent of the inputs and need not be introduced while estimating the frontier and thus the model can be re-written as

[Y.sub.it] = [[alpha].sub.i] + [X'.sub.it][beta] + [v.sub.it] (2)

The intercept [alpha] i is a MFI specific effect. The panel data method assumes that the error terms are uncorrelated with the independent variables. On satisfying this assumption, the MFI specific coefficients can be estimated using the Fixed Effect Method and the MFI specific technical efficiencies can be derived.

The level of a particular MFIs performance can be examined relative to the level achieved by the most efficient MFI. For 'n' MFIs there would be 'n' estimated performance measures given by [alpha]1, [alpha]2, [alpha]3, .... [alpha]n

If [alpha] = max ([alpha]n) is the performance of the most efficient MFI, the relative efficiency of the ith MFI would be

[Z.sub.i] = [[alpha].sub.i] - [alpha] (3)

Low value of [Z.sub.i] in absolute term implies that MFI 'i' is very inefficient relative to the most efficient MFI. The relative efficiency which is the deviation from that of the most efficient MFI is given as:

[S.sub.i] = [Z.sub.i] / [alpha] 1, 2, 3, .... n MFIs.

By assuming the most efficient MFI to be operating at 100 per cent efficiency (Schmidt and Sickles, 1984), the technical efficiency thereby computed allows to measure efficiency of the specific MFI relative to an absolute standard of 100 per cent.

VI. DATA SOURCES

The availability of Data on MFIs and its activities over the world and especially for India is very much limited. A couple of organization publishes secondary data regarding MFIs but the amount of information regarding them is very limited as of now, unlike the formal banking database. Therefore any analysis with the currently available secondary sources would have its own limitations. This study uses available information regarding twenty six sample MFIs in India. The data on these MFIs are taken from the website www.mixmarket.org, which is a website for microfinance information exchange.

The selection of sample was based on the availability of the required information for atleast latest three consecutive years, i.e., for the years 2005, 2006 and 2007. Total of around ninety MFIs in India were listed on the website but only twenty six of them had data for atleast the latest three years. These twenty six MFIs were a mix of NGOs, NBFCs and Credit Cooperative.

VII. INPUTS AND OUTPUTS

Considering financial institutions as decision making and service producing units the production approach function was followed to estimate the frontier and thereby the technical efficiency of MFIs in India. Under the production approach the financial institutions are considered as the producers of deposits and loans. The number of employees and capital expenditures are important inputs in this approach. The loans/ credit is the most important financial service that MFIs provides to their customers. Therefore this study selected loans disbursed by MFI as a single output. Main inputs required to produce loans are labour and Cost (Capital).

We have taken two inputs that are credit officers/staff as a proxy for labour and cost per borrower as a proxy for expenditures. Production approach suggests credit officers as input. The credit officers are relevant because they are actively engaged with loan portfolio of the MFIs.

On the whole the performance indicators can be classified into Outreach Indicators, Institutional Indicators and Financial Structure Indicators. Some of the outreach indicators are number of Active borrowers and Average Loan per borrower. On an average these numbers happen to be 109548 and 117.31.

Though the averages of these outreach indicators are impressive their standard deviations in Table I say that there is a huge variation in these indicators among different MFIs. In the financial indicators the profit margins, return on Assets, Gross Loan Portfolio, debt Equity ratio and operating expanses, all of them are having a high standard deviation and varies widely. Age of the MFIs also is having a remarkable seem to be an important characteristic of the MFIs that would determine the functioning and performance of the MFIs.

VIII. EFFICIENCY OF MFIS IN INDIA

The estimates from the stochastic production frontier model specified in equation (2) of MFIs in India are presented in the Table 2.

Whether to use fixed effects (FE) model or random effects (RE) model can be tested statistically using the Hausman test, which is a test of equality between the coefficients estimated via FE and RE models. Assuming that the model is correctly specified, a significant difference in the coefficient estimates is indicative of correlation between the individual effects and the repressors. Where this correction is present, the estimates using RE model will be biased (Vinish & Deepa, 2005).

The results of Hausman specification test statistic of (7.83) shown in Table 2. Suggests that the null hypothesis of no correlation is rejected and FE model is clearly preferable to RE model. The test statistic is significant at 5% Level. In the FE model the No. of loan officers (proxy for labour) is significantly affecting the dependent variable Average loan per borrower (output). The statistic is significant at 5% level. At the same time the other-variable cost per borrower (proxy for Capital) is statistically significant at 10% level.

IX. ANALYSIS OF EFFICIENCY DETERMINANTS

The average technical efficiency according to the estimates in Table 3. is 70%, the lowest being 48.8% for BANDHAN. The analysis reveals that Satin Credit care and IASC are the most efficient (the best practice) MFIs in India (2). The standard deviation of the technical efficiencies is 12.5 indicating that there is a substantial variation in technical efficiencies of various MFIs from the most efficient MFI.

This section investigates the possible determinants of efficiency of MFIs in India. Different variables that can explain the efficiency of MFIs has been proposed according to the theory. These variables can be divided into different groups based on the financial management and performance and the basic characteristics of MFIs. Regression analysis has been used to estimate the determinants of technical efficiency in this section.

First variable that has been considered is the characteristics of MFIs including age and size. The age may represent the experience of MFI. To capture the effect of the size of MFI the total value of assets (TA) was used. It is hypothesized that large with more experience MFIs may perform better than those having less experience and with smaller size.

The variable that represents the financial management of MFIs is Debt-Equity ratio. It is expected that higher debt-equity ratio reduces firms efficiency. The next set of variables represents the overall performance of the MFI, i.e. the return on assets (ROA) which is expected to have positive association with firm's efficiency. Lastly the individual nature of MFI has been taken into account to examine whether the NGO or NBFI status as such, influence the technical efficiency of the respective MFIs.

The OLS Regression Model that is used to estimate the determinants of Technical Efficiencies of the MFIs is be specified as:

TE = [alpha] + [[beta].sub.1](SIZE) + [[beta].sub.2](AGE) + [[beta].sub.3](TYPE) - [[beta].sub.4](DER) + [[beta].sub.5] (ROA) + [u.sub.i]

SIZE - Size of the MFI measures in terms of Total Assets

AGE = No. of years of Experience of MFI

TYPE = Status of MFI, whether NGO or NBFI; 1 if type = NGO, Else 0 (NBFI)

DER = Debt - Equity ratio

ROA = Return on Assets

The results of regression analysis are presented in Table 4. The value of adjusted R2 show that 65% of variation in the technical efficiency is explained by the variables included in the model. The parameter estimate of the size variable represented by the total value of assets is significant having positive sign. It implies that the size of the MFI is important in determining TE levels. At the same time the age in no. of years of experience has a positive and significant effect on TE. As the no. of years of experience for the MFI increase the TE also increases. In accordance with the theory, the debt-equity ratio has a inverse relationship with the TE meaning the more the debt the more the interest payment. The higher the interest payment, the higher the cost of capital and lesser the total output. The return on assets has appositive relationship with the MFI's technical efficiency but is not significantly affecting it.

X. CONCLUDING REMARKS

The objective of this study has been to benchmark the best practice MFIs thereby to estimate the efficiency and its determinants. For the efficiency analysis the Stochastic Production Frontier Approach is followed. The results from the regression analysis lead to conclude that size of MFI is important in the determination efficiency of MFIs. The second important conclusion is that the MFIs debt-equity ratio should be minimum, else it becomes less efficient. This is because the cost of funds increases and there by it reduces the potential output. So MFIs should borrow less and plough the profits inside. According to the technical estimates IASC, KBSLAB, ADARSHA & CMML are performing well having technical efficiency levels above 80%, Satin Credit Care being the frontier of 100% efficient or best practicing MFI. Further the analysis also reflects that, being an NBFI the chances to have better technical efficiency is more than being an NGO. So to conclude we can say that it is appropriate to benchmark the best practice MFIs at certain intervals in such a way that it would show the path to success to the other MFIs.

References

Begona Gutierrez-Nieto, Carlos Serrano-Cinca, Cecilio Mar Molinero (2007), "Microfinance Institutions and Efficiency", Omega, The International Journal of Management Science, Vol. 35, pp. 131-142.

Evans, D. B., Tandon, A., Murray, C. J. L. and Lauer, J. A. (2001), The Comparative Efficiency of National Health Systems in Producing Health: An Analysis of 191 Countries. GPE Discussion paper series No. 29. Geneva: WHO, EIP/GPE/EQC. http://www.sa-han.org/ CapacityBuilding/Building%20Sustainable%20MFIs%20in%20India.pdf

Leon, J. (2003), Exploring the Determinants of Cost-efficiency and Micro-finance institution Universidad Anabuac, Maxico.

M. J. Farrell (1957), The Measurement of Productive Efficiency, Journal of the Royal Statistical Society, Series A (General), Vol. 120, No. 3, pp. 253-290.

Marek Hudon (2006), "Subsidies and Financial Performances of the Microfinance Institutions: Does Management Matter?" Working Papers CEB from Universite Libre de Bruxelles, Solvay Business School, Centre Emile Bernheim (CEB).

Morduch, J. and B. Haley (2002), "Analysis of the Effects of Micro-finance on Poverty Reduction". New York: NYU Wagner Working Paper No. 1014.

Murray, C. J. L. and Frenk. J. (1999), "A WHO Framework of Health System Performance Assessment: Global Program on Evidence and Information for Policy", Washington. D.C, World Bank.

Shahid Khandker, Baqui Khalily (1996), "The Bangladesh Rural Advancement Committee's Credit Programs: Performance and Sustainability", World Bank Discussion Paper, No. 324, World Bank.

Rajaram Dasguptha (2001), "Working and Impact of Rural Self-help Groups and other Forms of Micro-Financing: An Informal Journey through Self-help Groups", Indian Journal of Agricultural Economics Vol. 56, No. 3, July-Sep.

Rhyne, Elisabeth (1998), "The Yin and Yan of Micro-finance: Reaching the Poor and Sustainability", Micro Banking Bulletin, Vol. 2(1), pp. 608.

Vijay Mahajan & G. Nagasri (1999), "Building Sustainable Microfinance Institutions in India", Proceedings of the Conference on New Development Finance, Frankfurt, September.

Vinish Kathuria and Deepa Sankar (2005), "Inter-State Disparities in Health Outcomes in Rural India: An Analysis Using a stochastic Production Frontier Approach", Development Policy Review, 23 (2): 145-163.

P. MAHENDRA VARMAN

Department of Econometrics, University of Madras, India

Notes

(1.) For references quoted "Nanda, NABARD" see Website http://www.gdrc.org/icm/nanda-link.html

(2.) The estimation of best practice MFI in this paper is among those for which data was available.

Table 1
Performance Indicators of MFIs in India

Indicators in Analysis Minimum Maximum Mean

Loan Officers (Personnel) (No.) 5 2456 414
Active Borrowers (No.) 426 972212 109548
Average Loan Balance per 18 482 117.31
 Borrower (in US $)
Gross Loan Portfolio (in US $) 28613 91683453 10933347.5
Total Assets (in US $) 37757 101488992 13479195.9
Debt Equity Ratio (%) 70.09 22620.03 1977.26
Return on Assets (%) -14.99 9.07 1.2601
Profit Margin (%) -13.11 52.78 5.488
Cost Per Borrower (in US $) 2 79 14.48
Operating Expense (in US $) 1668 9855194.15 1184622.46
Age (No. of Years) 0 17 10.54

Indicators in Analysis Standard
 Deviation

Loan Officers (Personnel) (No.) 587.93
Active Borrowers (No.) 201282.40
Average Loan Balance per 87.22
 Borrower (in US $)
Gross Loan Portfolio (in US $) 20201035.35
Total Assets (in US $) 23366896.53
Debt Equity Ratio (%) 3316.27
Return on Assets (%) 3.96
Profit Margin (%) 16.08
Cost Per Borrower (in US $) 14.60
Operating Expense (in US $) 2099842.05
Age (No. of Years) 3.65

Source: www.mixmarket.org

Table 2
Stochastic Production Frontier Estimates

Variable Fixed Effects Std. Random Effects Std. Error
 Coefficient Error Coefficients

Ln_Employees 0.1503 ** 0.632 0.267 0.375
Ln_C/Borrower 0.1860 * 0.118 0.377 * 0.732
[R.sup.2] 0.84 -- 0.44 --
Hausman Test 7.83 **, Prob. Value 2 (df) 0.0199

Dependent Variable: Ln_Avg-Loan/Borrower

Note: Ln_Employees [right arrow] Log of no. of Loan Officers Employed

Ln_C/Borrower [right arrow] Log of Cost per Borrower

Ln_Avg Loan/Borrower [right arrow] Log of Average loan outstanding per
borrower

(**) denotes significant 5 % Level , (*) denotes significant at
10 % level.

Table 3
Technical Efficiency Scores (%) and Ranks across MFIs

Name of MFIs MFI Coefficient [z.sub.i]- Technical
 ([z.sub.i]) deviation Efficiency (TE)

ADARSHA 3.88202 -0.73949 82.7
AMMACTS 3.65099 -0.97052 76.7
ASSIST 3.42988 -1.19163 70.7
Bandhan 2.69153 -1.92998 48.8
BASIX 3.40191 -1.2196 69.9
Bhoomika 3.32758 -1.29393 67.8
BISWA 2.91502 -1.70649 55.7
BSA 3.07409 -1.54742 60.4
CMML 3.78523 -0.83628 80.2
Coshpor MC 2.96927 -1.65224 57.3
CReSA 3.44328 -1.17823 71.0
GK 3.32897 -1.29254 67.8
GV 2.97464 -1.64687 57.5
IASC 4.57342 -0.04809 99.0
KBSLAB 3.89014 -0.73137 82.9
Kotalipara 2.9327 -1.68881 56.2
Mahasemam 3.10351 -1.518 61.3
NDFS 3.71721 -0.9043 78.4
PMS Indore 3.72621 -0.8953 78.6
Sanghamitra 3.32129 -1.30022 67.6
Sarvodaya nano finance 3.46019 -1.16132 71.5
Satin Credit care 4.62151 0 100.0
Share 3.01275 -1.60876 58.6
SKS 3.2422 -1.37931 65.3
SPADANA 3.22163 -1.39988 64.8
VWS 3.18532 -1.43619 63.7

Name of MFIs Rank (TE)

ADARSHA 4
AMMACTS 8
ASSIST 11
Bandhan 26
BASIX 12
Bhoomika 14
BISWA 25
BSA 20
CMML 5
Coshpor MC 23
CReSA 10
GK 13
GV 22
IASC 2
KBSLAB 3
Kotalipara 24
Mahasemam 19
NDFS 7
PMS Indore 6
Sanghamitra 15
Sarvodaya nano finance 9
Satin Credit care 1
Share 21
SKS 16
SPADANA 17
VWS 18

Table 4
Results of Regression analysis for Determinants of TE

Variables Coefficients t-statistic

Constant 70.864 *** 15.430
Size (Total Assets) 2.482 *** 3.008
Age (no. of Years of Experience) 0.813 *** 2.42
Type 1(NGO) -10.128 *** -3.11
ROA (% of Returns on Assets) 1.89 0.902
DER (% of Debt to Equity) -0.917 ** -1.82
[R.sup.2] 0.694
Adj. [R.sup.2] 0.652

Note: (***) denotes significant 1% Level,

(**) denotes significant at 5% level,

(*) denotes significant at 10% level